The FAO of the UN and the WHO (FAO/WHO) define ‘probiotics’ as ‘live micro-organisms which when administered in adequate amounts confer a health benefit to the host’(1). Lactic acid bacteria (LAB) and bifidobacteria constitute a significant proportion of probiotics, and many of the former are normal and non-pathogenic inhabitants of the human intestine. Among probiotics, Lactobacillus and Bifidobacterium are the most studied genera(Reference Kleerebezem and Vaughan2, Reference Lebeer, Vanderleyden and De Keersmaecker3). Members of these genera are active against gastrointestinal pathogens, such as Helicobacter pylori, rotavirus and urogenital pathogens(Reference Moreno Muñoz, Chenoll and Casinos4, Reference Juárez Tomás, Saralegui Duhart and De Gregorio5), in both in vitro and in vivo assays. In addition, some probiotic strains have been proven to be useful in the prevention of various disorders, including diarrhoea, allergy and inflammatory diseases(Reference Minocha6). Moreover, probiotics, as well as their metabolites, have been suggested to play an important role in the formation and establishment of a well-balanced intestinal microbiota in human newborns and adults(Reference He, Priebe and Zhong7, Reference Salazar, Ruas-Madiedo and Kolida8). Researchers have focused their attention on the isolation and characterisation of novel potential probiotic strains from different sources, primarily the gastrointestinal tracts of animals and human subjects, human milk and, less frequently, fruits and fermented foods(Reference Baruzzi, Poltronieri and Quero9–Reference Verdenelli, Ghelfi and Silvi13).
To be considered a probiotic, a strain should be able to colonise the gastrointestinal tract and promote host health through its metabolic activities. Specifically, probiotics should survive the acidic conditions of the stomach, resist relatively high levels of bile salts and adhere to the gut epithelium. Attempts to obtain a consensus and guidelines for probiotic safety evaluation have been made by the FAO/WHO(1) and the European Union (EU)-funded Product Safety Enforcement Forum of Europe (EU-PROSAFE) project(Reference Vankerckhoven, Huys and Vancanneyt14). These groups have recommended that the genus and species of the micro-organism must first be definitively determined by phenotypic and genotypic techniques and those strains should be deposited in an internationally recognised culture collection. For safety, at a minimum, the probiotic strains should also be characterised with respect to their antibiotic resistance patterns, certain metabolic activities (e.g. d-lactate production, bile salt deconjugation and biogenic amines), safety through the use of acute ingestion studies in murine models and the estimation of potential side effects during human studies. Similarly, because probiotics must confer a health benefit to the host(1), it is necessary to ascribe a functional role to the potential probiotic strain, such as evaluating its potential activity against the growth of a human pathogen(Reference Borchers, Selmi and Meyers15, Reference Dunne, O'Mahony and Murphy16).
The aim of the present study was to isolate, identify and characterise novel LAB strains with potential probiotic properties from the faeces of healthy, exclusively breast-fed infants. The isolated strains were selected primarily for their acid resistance, bile tolerance and ability to adhere to intestinal cells. Sequencing the 16S ribosomal RNA (rRNA) genes and 16S–23S rRNA intergenic spacer regions and assessing enzymatic properties then identified the isolated strains. Following the FAO/WHO guidelines, the isolated strains were deposited in an international type culture collection (Institut Pasteur). Furthermore, to ensure that the selected strains are not harmful to human health and confer a potential benefit to the host, a detailed toxicological characterisation, including metabolic activities, antibiotic resistance and acute ingestion assays, in murine models were performed. Also evaluations of their activity against human pathogens were conducted in Listeria monocytogenes and rotavirus. L. monocytogenes is an aetiological agent of meningitis and it is characterised by a high case-fatality rate and well-defined risk groups, including pregnant women, newborn infants and neonates(Reference Denny and McLauchlin17). Rotavirus is the leading cause of severe acute gastroenteritis among children worldwide(Reference Parashar, Gibson and Bresee18). Finally, a preliminary immunological study followed by a study on mice on the role of the isolated potential probiotics in enhancement of immunity was conducted.
Materials and methods
Subjects
A total of twelve healthy, exclusively breast-fed infants, aged 1 month, were selected for the study at the Clinic Hospital of the University of Granada. The present study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by the Ethical Committee of the University of Granada. Written informed consent was obtained from the parents after a careful explanation of the nature of the study.
Faecal samples and bacterial growth conditions
The fresh faecal samples were treated according to Thompson-Chagoyan et al. (Reference Thompson-Chagoyan, Vieites and Maldonado19). All samples were diluted and plated on Beerens medium(Reference Beerens20) and Bifidobacterium Medium (BFM) medium(Reference Nebra and Blanch21), which are both selective for Bifidobacterium spp., and on Rogosa medium (Oxoid), which is selective for Lactobacillus spp. The plates were incubated for 48–72 h at 37°C in an anaerobic or enriched CO2 atmosphere. For each faecal culture sample, 100 colonies (fifty lactobacilli and fifty bifidobacteria) were randomly selected and incubated in Man, Rogosa and Sharpe (MRS) broth medium (Oxoid) for Lactobacillus spp. and in MRS broth medium (Oxoid) supplemented with 0·05 % (w/v) cysteine (Sigma-Aldrich) (MRS plus cysteine (MRS-C) medium) for Bifidobacterium spp. After centrifugation, the cells were stored at − 80°C in glycerol until further analysis.
Characterisation of isolated lactic acid bacteria as potential probiotics
Resistance to low pH and bile salts
The resistance of the isolated bacteria to low pH and different concentrations of bile salts (Oxgall, Sigma-Aldrich) was evaluated by monitoring bacterial growth. Briefly, 900 μl of PBS, which was adjusted to pH 2·0, 2·5, 3·0 or 7·0 (control) or supplemented with 0·3, 0·5 or 0·7 (w/v) Oxgall (Sigma-Aldrich), was inoculated with 100 μl of a 48 h culture, previously washed three times with PBS. After of 3 h culture at 37°C under anaerobic conditions, 50 μl of the diluted cultures was spread on plates of selective culture media and incubated for 48–72 h at 37°C, followed by colony counting. The percentage of viable bacteria was calculated. The assays were performed in three independent experiments.
Only those strains that survived the resistance tests were assayed for adherence to intestinal cells; the commercial strains Lactobacillus rhamnosus GG American Type Culture Collection (ATCC) 53 103 and Bifidobacterium longum BB536 (Morinaga & Company Limited) were grown and used as controls for these assays.
Adhesion to HT-29 cells
The assay for bacterial adhesion to HT-29 cells was performed on the surviving strains described earlier. The cells were grown in twenty-four-well tissue culture plates (NUNC) in Dulbecco's modified Eagle's medium (DMEM); (Sigma-Aldrich) supplemented with 10 % (v/v) heat-inactivated fetal calf serum (Invitrogen), 2 mm-l-glutamine, 1 % (w/v) non-essential amino acid preparation (Sigma-Aldrich), 66 μg penicillin and 100 mg/ml streptomycin (Sigma-Aldrich). The cells were cultured at 37°C in an atmosphere of 5 % (v/v) CO2 and 95 % air until a confluent monolayer was reached.
Before the adhesion assay, the HT-29 cell monolayers were washed twice with antibiotic-free DMEM. Each well was inoculated with 250 μl of 48 h LAB culture, washed three times with PBS, diluted in DMEM to an optical density at 600 nm of 0·8 units (equivalent to 1010 bacteria) and incubated for 90 min at 37°C in a 5 % (v/v) CO2 atmosphere. After the incubation, each well was washed with PBS and 100 μl of trypsin-EDTA (Sigma-Aldrich) was added. An aliquot of 50 μl of the homogenate was spread on plates containing Lactobacillus spp.- or Bifidobacterium spp.-specific medium. The percentage of bacteria that adhered to the wells was calculated. The assays were performed in three independent experiments.
Identification and biochemical characterisation of isolated strains
The isolated strains that survived low pH, exhibited bile salt resistance and adhered well to intestinal cells were identified by sequencing the 16S rRNA gene and 16S–23S rRNA intergenic spacer region and characterised by enzymatic activities and carbohydrate utilisation. A DNA extraction procedure was performed using the phenol–chloroform method(Reference Sambrook, Fritsch and Maniatis22). The DNA was checked for purity using standard methods. The 16S rRNA from the selected strains was amplified by PCR with the universal primers 27f 1492r and 39f, as reported previously(Reference Klingberg, Axelsson and Naterstad23, Reference Haarman and Knol24). The 16S–23S intergenic spacer region was amplified with the primers LactoF (5′-ACACCGCCCGTCACACCATG-3′) and LactoR (5′-CCHSTTCGCTCGCCGCTACT-3′) for Lactobacillus (Reference Klingberg, Axelsson and Naterstad23) and the primers F_allbif_IS and R_allbif_IS for Bifidobacterium (Reference Haarman and Knol24). The amplification mixture (50 μl) contained 1 μl (1 nmol/μl) of each primer; 1 μl (1 U/μl) Taq DNA polymerase (Biotools B&M Labs); 5 μl 10 × reaction buffer (Biotools); 200 μm of each deoxynucleoside triphosphate (Biotools) and 1 μl DNA template. The DNA template was amplified by an initial incubation at 80°C for 5 min, denaturation at 94°C for 2 min, thirty-five cycles of denaturation at 94°C for 30 s, annealing at 55°C (27f-1492r), 65°C (39f-1391r) or 60°C (LactoR-LactoF; F_allbif_IS-R_allbif_IS) for 30 s and extension at 72°C for 1·5 min (27f-1492r; 39f-1391r) or 1 min (LactoR-LactoF; F_allbif_IS-R_allbif_IS), followed by a final extension at 72°C for 10 min. Controls devoid of DNA were included in the amplification process. The integrity of the PCR products was assayed by detection of single bands following electrophoresis. The PCR products were purified with the Illustra™ GFX™ PCR DNA and Band Purification Kit (General Electric Healthcare) and sequenced by the Genomic Service of the Institute of Parasitology and Biomedicine ‘Lopez-Neyra’ (Consejo Superior de Investigaciones Cientificas).
The resulting sequences were automatically aligned, inspected by eye and compared with the online tool BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi). The strains were identified based on the highest hit scores. The 16S rRNA sequences of L. paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and L. rhamnosus CNCM I-4036 were deposited in the National Center for Biotechnology Information (NCBI) nucleotide sequence database under the accession no. JQ621984, JQ621983 and JQ621982, respectively.
Carbohydrate fermentation and the enzymatic activities of each LAB strain were analysed with the API 50 CHL System Kit and API ZYM Kit, according to the manufacturer's protocol (BioMérieux).
Sensitivity to antibiotics
The sensitivity of each strain to twenty antibiotics was determined according to the European Food Safety Authority's (EFSA) recommendations(25). The minimum inhibitory concentration values were determined in LAB susceptibility test medium (LSM) broth formulation(Reference Klare, Konstabel and Müller-Bertling26) using the broth dilution antimicrobial susceptibility test established by the Clinical and Laboratory Standards Institute. The antibiotics were tested over a concentration range of 0·125–512 mg/l. The assays were performed in three independent experiments.
Toxicological study
Non-desired metabolite production
Lactic acid isomers production
Lactic acid production was determined in supernatants from a 17 h culture of each bacterial strain grown in their optimum conditions with a commercial kit (d-lactic acid/l-lactic acid; Megazyme). The assays were performed in three independent experiments.
Deconjugation of bile salts
Bile salt hydrolase (BSH) activity of intact and sonicated cells, as well as cell-free supernatants, was evaluated with the substrates glycocholate and taurocholate, according to the technique of Kumar et al. (Reference Kumar, Brannigan and Prabhune27). The assays were performed in three independent experiments.
Formation of biogenic amines
The formation of biogenic amines cadaverine, histamine, putrescine and tyramine in the cell-free supernatants of 17 h cultures was detected following the chromatographic method described by Eerola et al. (Reference Eerola, Hinkkanen and Lindfors28). The assays were performed in three independent experiments.
Acute ingestion study in immunocompetent and immunosuppressed mice
All procedures involving animals were conducted in accordance with the regulations established by the European Community Council on the protection of animals used in experimental and scientific applications (Regulation 86/609/EEC), and the experimental protocol was approved by the Biopolis Ethics Committee. Acute ingestion study was performed according to Chenoll et al. (Reference Chenoll, Casinos and Bataller29). In brief, the assays were performed in 7 week-old pathogen-free male BALB/c mice. Immunosuppressed group were achieved by intraperitoneal administration of cyclophosphamide (40 mg/kg per d), 5 d prior to the first bacterial administration and daily throughout the study, and kept inside containment units under positive pressure. After 2 weeks of acclimatisation, the mice were fed each day for 6 d with the three isolated strains (5·5 × 108 CFU) or placebo (lyophilised skin milk with sucrose (5 %, w/v)) in a total volume of 200 μl by oral gavage.
Mortality and morbidity were noted twice a day and individual body weights were recorded at the beginning and end of the trial. At day 7 of the study, blood from each mouse was collected by submandibular venipuncture and immediately the mice were killed by cervical dislocation. For the bacterial translocation assessment, liver, spleen and mesenteric lymph nodes were removed under aseptic conditions to avoid any cross-contamination and bacterial counts at 37°C on MRS agar for Lactobacillus strains and MRS-C for Bifidobacterium strains were obtained. For the histological examination, the duodenum–jejunum, the proximal colon, the distal ileum and the distal colon were immediately excised, fixed (10 % (v/v) buffered formalin) and embedded in paraffin. Pieces of 5 μm were stained with haematoxylin and eosin and examined by direct microscopy.
Inhibition of pathogens by lactobacilli and bifidobacteria supernatants
Obtaining the supernatants
To obtain supernatants, the bacterial strains were grown anaerobically for 17 h and 24 h at 37°C in MRS medium (lactobacilli) or MRS-C (bifidobacteria). After centrifugation at 12 000 g for 10 min, supernatants were neutralised to pH 6·5 with NaOH (1 m) and concentrated to 10 × by freeze-drying. Concentrated supernatants were sterilised by filtration through a 0·22 μm pore-size filter and stored at − 20°C until use.
Listeria monocytogenes strains and growth conditions
L. monocytogenes strains CECT 935, CECT 4031 and CECT 911 were obtained from the Spanish Type Culture Collection (CECT). L. monocytogenes was grown in Brain Heart Infusion (BHI) broth and incubated aerobically for 24 h at 37°C.
Cells lines and viruses
The human colon carcinoma cell line HT-29 was grown in DMEM supplemented with 10 % (v/v) fetal bovine serum (FBS). The human rotavirus Wa strain (TC adapted; ATCC VR-2018) was obtained from the ATCC. The human rotaviruses Ito and Va70 were kindly provided by Dr Javier Buesa (Hospital Clínico Universitario). The viral stocks of the human rotavirus strains were propagated by infecting HT-29 cells in the presence of 1 μg/ml trypsin (type IX; Sigma). Aliquots of the viruses were stored at − 80°C until use.
Activity of lactobacilli and bifidobacteria supernatants against Listeria monocytogenes
Assays were performed in polystyrene ninety-well (volume, 200 μl/well) multiwell plates (Maxisorp; Nunc). BHI broth was inoculated with a 5 % (v/v) concentrated L. monocytogenes cell solution, and the neutralised supernatant was added to a final concentration of 0·4, 2 or 4 % (v/v). The ability of the strains to inhibit L. monocytogenes was evaluated by monitoring bacterial growth at 37°C in BHI medium in ninety-six-well plates. Bacterial growth was analysed at 655 nm in a Multiskan microplate reader (Thermo Fisher Scientific). The percentage resistance was calculated in each case by comparing the final optical density with that of the corresponding controls, which were grown in BHI.
Rotavirus propagation and in vitro inhibition assays
Inhibition assays
Infection assays were carried out in HT-29 cell line according to Moreno Muñoz et al. (Reference Moreno Muñoz, Chenoll and Casinos4). Briefly, supernatants (1 × ) were added to HT-29 cell monolayers. After 1 h at 37°C, the supernatant was removed and the viral inoculum was added. After 1 h at 37°C DMEM medium was added and the plate was incubated for 15–18 h at 37°C with 5 % CO2. Prior to the immunoperoxidase assay, the infected HT-29 cell monolayers were fixed with methanol–acetone (1:1 (v/v)) for 15 min.
Immunoperoxidase assays
Following Moreno Muñoz et al. (Reference Moreno Muñoz, Chenoll and Casinos4), to detect viral antigens anti-VP6 monoclonal antibody 2F3E7 was added in fixed cells. The mixture was incubated at 37°C for 1 h and peroxidase-conjugated goat anti-mouse IgG antibody (Sigma) (100 μl/well, diluted 1/2000 in PBS-bovine serum albumin) was added and incubated at 37°C for 1 h. After washing with PBS, 100 μl/well of diamine-benzidine substrate was added. The infectious, peroxidase-stained foci were counted through an inverted microscope using five defined fields per well. The arithmetic mean was calculated to determine the number of foci per microscopic field, and these values were compared with the number of infectious foci in an untreated virus control.
Immunological studies
Experimental animal group
As described previously, all procedures involving animals were conducted in accordance with Regulation 86/609/EEC. All assay conditions (bacterial cell culture, housing of the mice, assigning the groups and killing the animal) were conducted as explained in the acute ingestion study. After 2 weeks of acclimatisation, the mice were fed with 200 μl of the three isolated strains or placebo by oral gavage every 48 h during 4 weeks. Mortality and morbidity were noted twice a day throughout the study. Individual body weights were recorded at the beginning and end of the trial.
Determination of total IgA and cytokine production
On the last day of the assay, faeces were collected. Determination of total IgA was performed as in Grewal et al. (Reference Grewal, Karlsen and Vetvik30). After killing, the spleen was removed under laminar flow and homogenised in PBS–bovine serum albumin buffer. The resulting homogenates were filtered through a cell strainer 70 μm (BD Falcon). Erythrocytes were lysed and splenocytes were inoculated in six-well plates at a concentration of 106 cells/well. Plates were incubated during 24 h at 37°C with 5 % CO2. Then, cultures were stimulated with ionomicine and PMA for 48 h. After this time, cell supernatants were recovered and cytokines interleukin (IL)2, IL-4, IL-10, IL-12 and interferon-γ were quantified by flow cytometry by using the FlowCytomix kit (Bender MedSystems).
Statistical analysis
The results are expressed as the mean and standard deviation. Differences between the mean values for different treatments with the isolated strains or their supernatants were analysed by one-way ANOVA. The least significant difference test was used for a posteriori r-paired comparison of the means. The statistical analysis was performed with Statgraphics plus (version 5.1) software (Manugistics).
For the murine assays, the organ weights were compared among the groups using Tukey's multiple comparison tests.
Results
Isolation and identification
A total of 4680 colonies were isolated from twelve faeces samples from breast-fed infants. Of these, 758 were resistant to low pH and were also tolerant to bile salts; these bacteria were tested for adhesion to HT-29 cells. In total, forty-two out of the 758 selected colonies strongly adhered to the cells, and consequently, their 16S rRNA genes were sequenced, resulting in the identification of three different strains that had greater than 99 % sequence identity to B. breve, L. paracasei and L. rhamnosus species. The three isolated strains were deposited in the Collection Nationale de Cultures de Microorganismes (CNCM) of Institute Pasteur. These strains were considered unique and were named as follows: L. paracasei CNCM I-4034, B. breve CNCM I-4035 and L. rhamnosus CNCM I-4036
Carbohydrate utilisation and enzymatic activities
An analysis of carbohydrate fermentation by the isolated LAB strains was done using the API 50 CHL System kit (see Materials and methods section for details). Both Lactobacillus strains utilised ribose, N-acetyl-glucosamine, arbutin, cellobiose, esculin, d-fructose, d-galactose, d-glucose, lactose, mannitol, d-mannose, melezitose, rhamnose, salicin, sorbitol, l-sorbose, trehalose, d-tagatose and d-turanose. Dulcitol, inulin and sucrose were utilised by the L. paracasei CNCM I-4034 strain but not by the L. rhamnosus CNCM I-4036 strain. Gentibiose, gluconate, maltose and α-methyl-d-glucoside were variably utilised by both strains, and adonitol, amygdalin and d-xylose were utilised only by L. paracasei CNCM I-4034 strain. The B. breve CNCM I-4035 strain utilised N-acetyl-glucosamine, amigdalin, l-arabinose, d-arabitol, l-arabitol, cellobiose, esculin, l-fucose, d-fructose, d-galactose, gentiobiose, gluconate, d-glucose, lactose, d-mannose, maltose, mannitol, melezitose, α-methyl-d-glucoside, d-raffinose, ribose, salicin, trehalose and d-turanose.
The enzymatic activities of the isolated LAB strains were also evaluated using the API ZYM kit (see Materials and methods section for details). L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036 strains exhibited acid phosphatase, alkaline phosphatase, arylamidase, cystine naphthol-AS-BI-phosphohydrolase, esterase (C4), esterase lipase (C8), β-galactosidase, α-glucosidase, leucine arylamidase and valine arylamidase activities. Only L. rhamnosus CNCM I-4036 strain exhibited α-chymotrypsin, α-fucosidase, α-galactosidase, β-glucuronidase and β-glucosidase activities. The strongest α-glucosidase, leucine arylamidase and valine arylamidase activities were exhibited by L. paracasei CNCM I-4034 strain. The strongest acid phosphatase, alkaline phosphatase, esterase (C4), esterase lipase (C8), α-fucosidase, β-glucosidase, naphthol-AS-BI-phosphohydrolase and valine arylamidase activities were exhibited by L. rhamnosus CNCM I-4036 strain. The B. breve CNCM I-4035 strain exhibited acid phosphatase, esterase (C4), esterase lipase (C8), α-galactosidase, β-galactosidase, α-glucosidase, β-glucosidase, leucine arylamidase, lipase (C14) and naphthol-AS-BI-phosphohydrolase activities. In particular, α-galactosidase, β-galactosidase and α-glucosidase activities were very strong in this strain.
Resistance to low pH, tolerance to bile salts and adhesion studies
The in vitro resistance of the isolated LAB strains to low pH and bile salts (using Oxgall, Sigma-Aldrich) were compared with that of two commercial strains (L. rhamnosus GG and B. longum from Morinaga & Company Limited (see results in Table 1). The five strains were able to grow in the presence of bile salts. The B. breve CNCM I-4035 and the L. paracasei CNCM I-4034 strain, as well as the two commercial strains, survived at pH 3·0. From all the tested strains, L. rhamnosus CNCM I-4036 was the most resistant to low pH, with 76·2 % survival at pH 2·0. Also the results of the adhesion of the isolated LAB strains to HT-29 cells are shown in Table 1. Results indicate that isolated LAB strains adhered to HT-29 cells more efficiently than the two commercial analysed strains.
Resistance to pH and bile salts and adhesion to HT-29 cells of the Lactobacillus paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and Lactobacillus rhamnosus CNCM I-4036 strains, and the control strains L. rhamnosus GG and Bifidobacterium longum
CNCM, Collection Nationale de Cultures de Microorganismes.
Sensitivity to antibiotics
The isolated strains did not show resistance to antibiotics (results not shown), except for 256 μg/ml of metronidazole, nalidixic acid and sulphonamide, and both the isolated Lactobacillus strains were resistant to 256 μg/ml of vancomycin.
Toxicological evaluation
Ex vivo assays
The results of lactic acid production assays are shown in Table 2. The B. breve CNCM I-4035 strain produced lower levels of l-lactic acid than the commercial strain B. longum Morinaga. d-Lactic acid levels were very low in both bifidobacteria. L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036 produced slightly more l-lactic acid than Lactobacillus GG strain, whereas the levels of d-lactic acid produced by both strains were similar to that obtained from the commercial strain. There was no detectable BSH activity with either taurocholate or glycocholate in the supernatants, intact cells or sonicated cells from 17 h cultures of L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036. BSH activity was observed in the supernatant from 17 h cultures and sonicated cells of B. breve CNCM I-4035 in the presence of glycocholate. Of the commercial strains, only the supernatant of Lactobacillus GG exhibited BSH activity in the presence of glycocholate, and the BSH activity of the B. longum Morinaga strain was higher than that of the B. breve CNCM I-4035 strain (Table 2).
Detection of undesirable metabolites in both cells and supernatants of the Lactobacillus paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and Lactobacillus rhamnosus CNCM I-4036 strains, and the control strains L. rhamnosus GG and Bifidobacterium longum (Mean values and standard deviations, n 3)
CNCM, Collection Nationale de Cultures de Microorganismes; BSH, bile salt hydrolase,
The results of the biogenic amine quantification in cell-free supernatants are summarised in Table 2. Putrescine was not detected from any of the isolated strains. For other amines, the values ranged from 0·70 to 4·85 for L. rhamnosus CNCM I-4036, 0·60 to 5·00 for L. paracasei CNCM I-4034 and 0·60 to 6·70 for B. breve CNCM I-4035. The values of these three amines for the B. longum strain were lower than those obtained for B. breve CNCM I-4035. For Lactobacillus GG strain, the values of cadaverine and histamine were lower than those obtained for L. paracasei CNCM I-4034 and L. rhamnosus CNCM I-4036. Of the considered amines, tyramine levels were the highest for the three isolated strains.
In vivo assays
Neither mortality nor adverse clinical signs were observed during the study. There were no statistically significant differences in body weight gain between the placebo and the treated groups (Table 3). In the immunosuppressed groups, there was a statistically significant loss in body weight throughout the study. None of the bacterial counts of the three isolated strains were found in the blood, spleen and liver. There were slight, but not significant, difference in the duodenum/jejunum and the ileum villus height and depth of the colonic crypts of the proximal and distal colon in the immunocompetent and immunosuppressed mice (data not shown). The Lieberkühn crypts had a uniform and constant aspect. However, the Brunner and Lieberkühn glands exhibited hyperplasia in the duodenum of the immunosuppressed mice, an effect that was not observed in immunocompetent mice. In the immunosuppressed mice treated with placebo, hyperplasia appeared in all cases (Table 3).
Results of acute ingestion assays of the Lactobacillus paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and Lactobacillus rhamnosus CNCM I-4036 strains, and the control strains L. rhamnosus GG and Bifidobacterium longum in immunocompetent and immunosuppressed mice† (Mean values and standard deviations)
CNCM, Collection Nationale de Cultures de Microorganismes; IC, immunocompetent mice; IS, immunosuppressed mice.
Values were significantly different: * P< 0·05; *** P< 0·001.
† Body weight gain is expressed as the final weight minus the initial weight.
‡ Data are expressed as the means and standard deviations of six doses.
§ Total germinal centres.
The three isolated strains yielded slight but not significantly different numbers of goblet cells in the small intestine. In the immunocompetent mice inoculated with the three isolated strains, the lymphoid tissue associated with the intestinal mucosa exhibited a higher number of lymphatic follicles and lower number of germinal centres than in the immunosuppressed mice. In both the immunosuppressed and immunocompetent mice, there were no statistically significant differences among groups treated with different isolated strains. The L. paracasei CNCM I-4034 and placebo treatment groups tended to have a higher number of lymphatic follicles. The L. rhamnosus CNCM I-4036-treated groups exhibited the lowest number of lymphatic follicles in both the immunosuppressed and immunocompetent groups. The immunosuppressed mice had more germinal centres than the immunocompetent mice. There were no statistically significant differences among the different treatments in the immunocompetent and immunosuppressed mice (Table 3).
Pathogen inhibition assays
Assay of activity against L. monocytogenes
The activity of the analysed supernatants against the three different strains of L. monocytogenes was variable and specific to the strain (Table 4). The 17 h supernatant of L. paracasei CNCM I-4034 primarily inhibited two L. monocytogenes strains (CECT 935 and CECT 4031). The 24 h supernatant of L. rhamnosus CNCM I-4036 at 2 and 4 % inhibited all of the L. monocytogenes strains. The L. monocytogenes CECT 911 strain was inhibited by both Lactobacillus strains at all supernatant percentages. The supernatant of B. breve inhibited L. monocytogenes CECT 4031 regardless of the supernatant percentage, and the less-concentrated supernatant (0·4 % (v/v)) inhibited L. monocytogenes CECT 911.
In vitro activity of Lactobacillus paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and Lactobacillus rhamnosus CNCM I-4036 strain supernatants from 17 and 24 h cultures (0·4, 2 and 4 % (v/v)) against Listeria monocytogenes strains (Mean values and standard deviations, n 3)
CNCM, Collection Nationale de Cultures de Microorganismes.
*P< 0·05; ** P< 0·01; *** P< 0·001.
Rotavirus inhibition assays
HT-29 cells infected with the human rotavirus strains Ito, Va70 and Wa were used in inhibition assay using 1 × supernatants from 17 and 24 h cultures (Table 5) . All three viruses were inhibited by 24 h supernatants of L. rhamnosus CNCM I-4036. L. paracasei CNCM I-4034 inhibited only the Wa and Va70 virus strains, whereas the B. breve supernatant did not inhibit any of the viruses.
In vitro activity of Lactobacillus paracasei CNCM I-4034, Bifidobacterium breve CNCM I-4035 and Lactobacillus rhamnosus CNCM I-4036 strain supernatants from 17 and 24 h cultures against the human rotaviruses Ito, Va70 and Wa in HT-29 cells (Mean values and standard deviations, n 3)
CNCM, Collection Nationale de Cultures de Microorganismes.
*P< 0·05.
Immunological effects
Also, in this case, there were no statistically significant differences in body weight gain between the placebo and treated groups (data not shown). The measurement of IgA concentration in faeces was higher in the case of the group fed with the B. breve CNCM I-4035 strain (results not shown). No statistical differences were detected between the four experimental groups for the measurement of most of the different assayed cytokines (IL-2, IL-4, IL-12 and interferon-γ). Only in the case of IL-10, there was a higher production by the B. breve CNCM I-4035 strain (results not shown).
Discussion
The beneficial properties of probiotics and the increased human consumption of these products have augmented efforts to identify potential probiotic strains. Selection and identification criteria for probiotic strains are now considered essential. In selecting potential probiotic strains, species identification by 16S rRNA gene and 16S–23S intergenic spacer region sequence analyses and the evaluation of their physico-chemical, safety and functional properties, such as resistance to gastric acid and tolerance to bile salts, are highly important(1).
There are many molecular tools for the identification of micro-organism species. Among them, 16s rRNA gene sequencing is the most frequently used, because it is extremely useful for determining phylogenetic relationships among organisms from the level of domains to the level of moderately closely related species(Reference Lane, Stackebrandt and Goodfellow31, Reference Stackerbrandt and Goebel32). By comparing the 16S ribosomal DNA sequences of the isolated strains with the sequences available in NCBI/BLAST (100 % homology), the isolated strains were identified as L. paracasei, L. rhamnosus and B. breve species, respectively. These bacteria are known to be present in the faeces of breast-fed infants, in which the genus Bifidobacterium accounts for 40–60 % of the total microbiota and B. breve species are present in a high percentage(Reference Harmsen, Wildeboer-Veloo and Raangs33). Additionally, a relatively high percentage of Lactobacillus, mainly L. casei, L. paracasei and Lactobacillus acidophilus, has been described(Reference Heiling, Zoetendal and Vaughan34, Reference Satokari, Vaughan and Favier35).
In the present study, the API 50CH fermentation system was used to test the carbohydrate fermentation ability of the isolated LAB strains. This system permits the metabolic characterisation of strains on a wide range of individual substrates, an important component of lactobacilli characterisation. The API ZYM system allows strains to be characterised with respect to enzymatic type and activity level. In the analysis of carbohydrate utilisation, the three probiotic strains exhibited high β-galactosidase activity, which is extremely relevant to the utilisation of lactose and could potentially serve to alleviate lactose intolerance in human subjects. Strains B. breve CNCM I-4035 and L. rhamnosus CNCM I-4036 exhibited high α-galactosidase activity, which is important for the hydrolysis of α-d-galactosyl-oligosaccharides, which are found in relatively high amounts in human milk (1 g/l)(Reference Ninonuevo, Park and Yin36) and permit the selective growth of Bifidobacterium in the intestine(Reference Gopal, Sullivan and Smart37). Of particular interest is β-glucuronidase activity, a known carcinogenic enzyme. It was not detected in L. paracasei CNCM I-4034, B. breve CNCM I-4035 or L. rhamnosus CNCM I-4036. In general, we can state that the three isolated probiotic strains exhibited low enzymatic activities on mannose, fucose and glucuronides, instead preferring carbon sources like lactose and glucose. However, L. rhamnosus CNCM I-4036 showed high activity for α-fucosidase, and B. breve metabolised l-fucose. Several neutral oligosaccharides in human milk contain fucose. In fact, the most abundant human milk oligosaccharide is 2-fucosyl-lactose(Reference Ninonuevo, Park and Yin36, Reference Newburg, Ruiz-Palacios and Altaye38), and these strains could contribute to the metabolism of human milk oligosaccharides in infants.
The carbohydrate fermentation assays demonstrated that the isolated Lactobacillus strains have similar profiles as the commercial L. rhamnosus GG control and other Lactobacillus strains(Reference Charteris, Kelly and Morelli39). These strains fermented l-rhamnose, an activity found in L. rhamnosus and L. paracasei strains. Interestingly, the L. paracasei CNCM I-4034 strain utilised inulin, an activity that was absent in the commercial L. rhamnosus GG control. This polysaccharide is classified as a prebiotic and is used in functional foods because it has been reported to increase the prevalence of Bifidobacterium in the colon(Reference Roberfroid, Van Loo and Gibson40–Reference Makras, Van Acker and De Vuyst43). Although some Lactobacillus species have been reported to grow in the presence of this prebiotic(Reference Makras, Van Acker and De Vuyst43, Reference Cebeci and Gurakan44), this trait is quite uncommon in this genus.
Probiotics must be able to survive the passage through the upper digestive tract to reach the large intestine(Reference Bezkorovainy45, Reference Tuomola, Crittenden and Playne46); thus, the isolated strains were exposed for 3 h to pH 3·0, 2·5 or 2·0 to select strains that were most resistant to a low pH environment. The in vitro survival test revealed that the three strains were resistant to pH 3·0 even after 3 h of exposure, although 2 h may be sufficient for passage through the stomach. However, only the L. rhamnosus CNCM I-4036 strain exhibited resistance to pH 2·5 and 2·0. The resistance of probiotics to low pH varies greatly depending on the species and strain. In general, Lactobacillus strains exhibit better resistance to low pH than Bifidobacterium strains(Reference Takahashi, Xiao and Miyaji47, Reference Matsumoto, Ohishi and Benno48). Resistance to low pH is an important characteristic for the food industry, especially in the design of functional foods, because probiotics will be able to withstand acidic environments for long periods, such as in fermented foods in which post-acidity could affect strain viability(Reference Jayamanne and Adams49).
Bile salts play a fundamental role in the defence mechanism of the gut. The relevant physiological concentrations of human bile range from 0·3 to 0·5 %(Reference Zavaglia, Kociubinsky and Pérez50, Reference Mainville, Arcand and Farnworth51). For Bifidobacterium and Lactobacillus, the resistance to bile salts varies between species and strains(Reference Ruas-Madiedo, Hernández-Barranco and Margolles52). In the present study, the three isolated strains and commercial controls exhibited a high tolerance to bile salts, up to 0·7 % (w/v). The adaptation to bile salts is related to changes in carbohydrate fermentation, glycosidase activity(Reference Burns, Sánchez and Vinderola53), exopolysaccharide production(Reference Ruas-Madiedo, Gueimonde and Arigoni54, Reference Ruiz, Sánchez and Ruas-Madiedo55), the composition of membrane proteins and fatty acids(Reference Gueimonde, Margolles and de los Reyes-Gavilán56), increased adhesion to human mucus and inhibition of pathogen adhesion(Reference Gueimonde, Noriega and Margolles57, Reference Alander, Satokari and Korpela58).
An important characteristic of probiotics within the intestinal microbiota is their capacity for adhesion to the intestinal epithelium; this avoids their elimination by peristaltic movement. Additionally, adhesion is a prerequisite for colonisation(Reference Forestier, de Champs and Valtoux59) and is a factor in the competitive exclusion of enteropathogens(Reference Lee, Puong and Ouwehand60), stimulation of the immune system(Reference Schiffrin, Rochat and Link-Amster61) and antagonistic activity against enteropathogens(Reference Coconnier, Bernet and Kerne'is62). The reported adhesion ability of Bifidobacterium and Lactobacillus strains varies depending on the in vitro method utilised(Reference Laparra and Sanz63). In the present study, HT-29 cells were used to study the adhesion ability of the isolated strains. The commercial control, the Lactobacillus GG strain, exhibited half (4 %) of the adhesion level of the commercial control B. longum (8 %), and the three isolated strains exhibited almost twice the level of in vitro adhesion of the commercial controls. Although the results of the in vitro studies cannot be directly applied to the in vivo situation, there is evidence to support an association between them(Reference Crociani, Grill and Huppert64).
Once the three potential probiotic strains were isolated, identified and characterised, their safety was assessed. Although these species belong to the list of taxonomic units proposed for qualified presumption of safety status(65) and thus considered innocuous, a detailed toxicological study was performed following the FAO/WHO recommendations(1). The resistance to antibiotics was similar to that previously reported(Reference Klare, Konstabel and Werner66). The strains for which the minimum inhibitory concentration was above the breakpoints recommended by the European Food Safety Authority(25) require further investigation. Whole-genome sequencing of the three strains would be useful to ensure the absence of antibiotic resistance genes in plasmids or between mobile genetic elements. The European Food Safety Authority has not published upper levels for d-lactic acid isomer and BSH activity, and comparisons with commercial probiotic strains could be a good approach to evaluating the safety of novel strains with respect to undesirable metabolite production. The level of production of both d- and l-lactic acid isomers was similar in the commercial L. rhamnosus GG strain and the isolated lactobacilli. Moreover, the level of d-lactic acid production in the B. breve strain was similar to that in the B. longum Morinaga strain, but l-lactic acid production was lower. The B. breve CNCM I-4035 strain displayed BSH activity with glycocholate in the supernatant and the sonicated cells, but the activity was lower than that of the commercial B. longum strain and that previously reported for other bifidobacteria(Reference Tanaka, Hashiba and Kok67). The lactobacilli strains did not display BSH activity, which is consistent with previous studies in which only certain strains of lactobacilli exhibited this activity(Reference Begley, Hill and Gahan68).
Biogenic amine levels were determined from supernatants; putrescine was not detected in any strain, and cadaverine, histamine and tyramine levels were negligible compared with the maximums recommended by the FAO/WHO. The in vivo acute ingestion study demonstrated a lack of mortality and morbidity after the inoculation of mice with the isolated strains, even in immunosuppressed mice. Moreover, the administration of the three strains did not lead to bacterial loads in organs or changes in histomorphology.
To study their capacity to inhibit the growth of pathogens, we tested the isolated strains against bacteria and viruses. L. monocytogenes is a Gram-positive pathogen that serves as an important model for understanding host immune resistance to intracellular bacteria. This intracellular bacterium infects human subjects through the ingestion of contaminated food and, in risk groups, including neonates, can cause meningitis with a high case-fatality rate(Reference Denny and McLauchlin17). The isolated strains inhibited L. monocytogenes in a strain-specific manner. L. rhamnosus CNCM I-4036 inhibited the three L. monocytogenes used in the present study, whereas L. paracasei CNCM I-4034 and B. breve CNCM I-4035 inhibited at least one strain of L. monocytogenes. Other authors have reported the capacity of some LAB to protect against experimental listeriosis(Reference Bambirra, Lima and Franco69–Reference Sato71).
Rotavirus infects mature enterocytes of the intestinal villus, and consequently, crypt cells are spared(Reference Greenberg and Estes72). Rotaviral diarrhoea has been attributed to different mechanisms and accounts for an estimated 2 million hospitalisations per year(Reference Parashar, Gibson and Bresee18). The probiotics isolated in the present study decreased the propagation of three rotavirus strains in an in vitro assay. These results agree with those of other authors(Reference Moreno Muñoz, Chenoll and Casinos4, Reference Greenberg and Estes72), who reported that probiotics can produce bioactive compounds that can modulate the protection of epithelial cells from rotavirus infection. Probiotics have been used to prevent some intestinal pathogenic infections, such as Salmonella, Shigella, Escherichia coli, Listeria and H. pylori (Reference de LeBlanc Ade, Castillo and Perdigon73–Reference Tsai, Huang and Lin77). The possible mechanisms of action for this protection include the production of acid(Reference Makras and De Vuyst78, Reference Midolo, Lambert and Hull79) and other by-products of bacterial metabolism(Reference Ghalfi, Allaoui and Destain76, Reference Kuleasan and Cakmakci80, Reference Olasupo, Olukoya and Odunfa81).
Finally, a preliminary study on the role of the isolated potential probiotics in enhancement of immunity was conducted. Results indicated that B. breve CNCM I-4035 strain has some properties of interest that needs to be confirmed in future experiments.
In conclusion, the results presented here demonstrate the capacity of the strains L. paracasei CNCM I-4034, B. breve CNCM I-4035 and L. rhamnosus CNCM I-4036, which were isolated from the faeces of exclusively breast-fed infants, to resist low pH and high bile salt concentrations and adhere to the colon. In addition, these novel strains inhibit some strains of L. monocytogenes and the infection of human cells with rotavirus in vitro. Also, the B. breve CNCM I-4035 strain has some promising immunological properties. Furthermore, toxicological studies demonstrated that these strains fulfil the main criteria required for safe human consumption. These strains could be useful in the prevention and treatment of diarrhoea due to rotavirus in infancy and in the prevention of meningitis mediated by L. monocytogenes. Further studies are needed to evaluate the in vitro and in vivo activities of the isolated strains against different enteropathogens, and human clinical trials must be performed before these strains can be commercialised.
Acknowledgements
S. M.-Q. and E. C. carried out the majority of the experiments and were responsible to write the first draft of the manuscript; they contributed equally to the present work. J. M. V. and A. S. were involved in the isolation and characterisation of bacterial strains. J. M. selected the infants and was responsible for obtaining the faecal sample and E. M., M. J. B. and F. R. participated in the process of strain characterisation. E. C., S. G. and D. R. were involved in the safety and L. monocytogenes and human rotavirus infection. M. B.-B. and C. G.-L. helped in the bacterial characterisation and statistical analysis. D. R. and A. G. were the leader scientists of the present work and contribute equally to the design and supervision of the experiments and results; they also were involved in writing the manuscript. The present work was supported by the HERO Group through its company HERO Spain S. A (contract no. 3143 signed with the Fundación General Universidad de Granada Empresa contract and a private contract signed with the Spanish Biotechnology Company Biópolis S. L.). Hero Spain S. A in turn was funded by the CDTI, Spanish Ministry of Health. C. G.-L. is a recipient of a postdoctoral fellowship from Plan Propio of the University of Granada. E. M., M. J. B. and F. R., are members of the Hero Global Technology Centre for Infant Nutrition. This centre forms part of the food company HERO with headquarters in Switzerland. E. C., S. G. and D. R. are members of Biópolis S. L. (Spain), a spin-off of the High Scientific Research Council (Consejo Superior de Investigaciones Cientificas), Ministry of Education, Spain. No other authors declare conflict of interest.
